Airborne Escherichia coli generated from poultry litter concentrated on coarse particles. Of the viable organisms recovered, 98.89 percent occupied particles larger than 2.1 micron, while 57.60 percent of the dust mass itself was below 1.0 micron. The half life of the airborne organisms was 5.7 plus or minus 1.2 minutes. The authors attribute the mismatch between where the dust sits and where the organisms sit to a shielding effect: organisms on fine particles die faster, so the recoverable fraction concentrates on coarser carriers even when the dust skews fine.
That mechanism is the paper's contribution to the carrier particle argument. It supplies a stated reason, rather than an assumption, for expecting viable organisms to be found on larger carriers than a dust size distribution alone would suggest. It runs parallel to the finding in Alonso, Raynor, Davies and Torremorell 2015 that virus viability in air is particle size dependent, arriving at a compatible conclusion from bacterial survivability rather than from viral isolation.
Two qualifications belong with any use of this work. It is a BSL-2 chamber study using aerosolized E. coli ATCC 25922 mixed with litter collected from a commercial farm, not a barn field study, and it should never be cited as field evidence. The 2.1 micron threshold is the Andersen cascade impactor stage cut point, so it marks where the instrument divides its stages and not where biology divides. Read within those limits, the shielding effect is the stated basis for biasing inferred carrier ranges toward the coarser ASHRAE bands when direct measurement is unavailable.